A Front End Expert System for the Calibration of SWMM Runoff Block

by J. W. Delleur, Purdue Univ, United States,
C. Baffaut, Purdue Univ, United States,



Document Type: Proceeding Paper

Part of: Critical Water Issues and Computer Applications

Abstract:

An expert system is discussed which acts as a front end to the Storm Water Management Model (SWMM) for the purpose of preparation of data input and parameter selection. The structure of knowledge representation used in this expert system is based on production rules. These production rules are cast in the form of if...then pairs similar to the logic implication statement 'if this condition holds, then this conclusion or action is appropriate'. The inference procedure used is backward chaining which works from the hypotheses to the facts. The most important part of the calibration SWMM occurs in the RUNOFF block which generates the hydrograph of each subcatchment. In addition the runoff block includes the choice of computation options and the estimation of the parameters which most uncertain and difficult to evaluate. To achieve the calibration the expert system divides the task into three components: 1) the initial estimation of the parameters, 2) the calibration diagnosis, and 3) the adjustment of the parameter values.

Subject Headings: Runoff | Expert systems | Calibration | Stormwater management | Parameters (statistics) | Artificial intelligence (AI) | Hydrographs

Services: Buy this book/Buy this article

 

Return to search